Electron Playwright Helpers

Helper functions to make it easier to use Playwright for end-to-end testing with Electron. Parse packaged Electron projects so you can run tests on them. Click Electron menu items, send IPC messages, get menu structures, stub dialog.showOpenDialog() results, etc.

Installation

npm i -D electron-playwright-helpers

Usage

For a full example of how to use this library, see the electron-playwright-example project. But here's a quick example:

Javascript:

const eph = require('electron-playwright-helpers')
// - or cherry pick -
const { findLatestBuild, parseElectronApp, clickMenuItemById } = require('electron-playwright-helpers')

let electronApp: ElectronApplication

test.beforeAll(async () => {
  // find the latest build in the out directory
  const latestBuild = findLatestBuild()
  // parse the packaged Electron app and find paths and other info
  const appInfo = parseElectronApp(latestBuild)
  electronApp = await electron.launch({
    args: [appInfo.main], // main file from package.json
    executablePath: appInfo.executable // path to the Electron executable
  })
})

test.afterAll(async () => {
  await electronApp.close()
})

test('open a file', async () => {
  // stub electron dialog so dialog.showOpenDialog() 
  // will return a file path without opening a dialog
  await eph.stubDialog(electronApp, 'showOpenDialog', { filePaths: ['/path/to/file'] })

  // call the click method of menu item in the Electron app's application menu
  await eph.clickMenuItemById(electronApp, 'open-file')

  // get the result of an ipcMain.handle() function
  const result = await eph.ipcMainInvokeHandler(electronApp, 'get-open-file-path')
  
  // result should be the file path
  expect(result).toBe('/path/to/file')
})

Typescript:

import * as eph from 'electron-playwright-helpers'
// - or cherry pick -
import { electronWaitForFunction, ipcMainCallFirstListener, clickMenuItemById } from 'electron-playwright-helpers'

// then same as Javascript above

Contributing

Yes, please! Pull requests are always welcome. Feel free to add or suggest new features, fix bugs, etc.

Please use Conventional Commit messages for your commits. This project uses semantic-release to automatically publish new versions to NPM. The commit messages are used to determine the version number and changelog. We're also using Prettier as our code format and ESlint to enforce formatting, so please make sure your code is formatted before submitting a PR.

Additional Resources

• Electron Playwright Example - an example of how to use this library
• Playwright Electron Class - Playwright API docs for Electron
• Electron API - Electron API documentation

API

Functions

findLatestBuild(buildDirectory) ⇒ string

Parses the out directory to find the latest build of your Electron project. Use npm run package (or similar) to build your app prior to testing.

Assumptions: We assume that your build will be in the out directory, and that the build directory will be named with a hyphen-delimited platform name, e.g. out/my-app-win-x64. If your build directory is not out, you can pass the name of the directory as the buildDirectory parameter. If your build directory is not named with a hyphen-delimited platform name, this function will not work. However, you can pass the build path into parseElectronApp() directly.

parseElectronApp(buildDir) ⇒ ElectronAppInfo

Given a directory containing an Electron app build, or the path to the app itself (directory on Mac, executable on Windows), return a bunch of metadata, including the path to the app's executable and the path to the app's main file.

Format of the data returned is an object with the following properties:

• executable: path to the app's executable file
• main: path to the app's main (JS) file
• name: name of the app
• resourcesDir: path to the app's resources directory
• asar: true if the app is using asar
• platform: OS platform
• arch: architecture
• packageJson: the JSON.parse()'d contents of the package.json file.

electronWaitForFunction(electronApp, fn, arg) ⇒ Promise.<void>

Wait for a function to evaluate to true in the main Electron process. This really should be part of the Playwright API, but it's not.

This function is to electronApp.evaluate() as page.waitForFunction() is page.evaluate().

stubDialog(app, method, value) ⇒ Promise.<void>

Stub a single dialog method. This is a convenience function that calls stubMultipleDialogs for a single method.

Playwright does not have a way to interact with Electron dialog windows, so this function allows you to substitute the dialog module's methods during your tests. By stubbing the dialog module, your Electron application will not display any dialog windows, and you can control the return value of the dialog methods. You're basically saying "when my application calls dialog.showOpenDialog, return this value instead". This allows you to test your application's behavior when the user selects a file, or cancels the dialog, etc.

Note: Each dialog method can only be stubbed with one value at a time, so you will want to call stubDialog before each time that you expect your application to call the dialog method.

stubMultipleDialogs(app, mocks) ⇒ Promise.<void>

Stub methods of the Electron dialog module.

Playwright does not have a way to interact with Electron dialog windows, so this function allows you to mock the dialog module's methods during your tests. By mocking the dialog module, your Electron application will not display any dialog windows, and you can control the return value of the dialog methods. You're basically saying "when my application calls dialog.showOpenDialog, return this value instead". This allows you to test your application's behavior when the user selects a file, or cancels the dialog, etc.

stubAllDialogs(app) ⇒ Promise.<void>

Stub all dialog methods. This is a convenience function that calls stubMultipleDialogs for all dialog methods. This is useful if you want to ensure that dialogs are not displayed during your tests. However, you may want to use stubDialog or stubMultipleDialogs to control the return value of specific dialog methods (e.g. showOpenDialog) during your tests.

ipcMainEmit(electronApp, message, ...args) ⇒ Promise.<boolean>

Emit an ipcMain message from the main process. This will trigger all ipcMain listeners for the message.

This does not transfer data between main and renderer processes. It simply emits an event in the main process.

ipcMainCallFirstListener(electronApp, message, ...args) ⇒ Promise.<unknown>

Call the first listener for a given ipcMain message in the main process and return its result.

NOTE: ipcMain listeners usually don't return a value, but we're using this to retrieve test data from the main process.

Generally, it's probably better to use ipcMainInvokeHandler() instead.

ipcMainInvokeHandler(electronApp, message, ...args) ⇒ Promise.<unknown>

Get the return value of an ipcMain.handle() function

ipcRendererSend(page, channel, ...args) ⇒ Promise.<unknown>

Send an ipcRenderer.send() (to main process) from a given window.

Note: nodeIntegration must be true and contextIsolation must be false in the webPreferences for this BrowserWindow.

ipcRendererInvoke(page, message, ...args) ⇒ Promise.<unknown>

Send an ipcRenderer.invoke() from a given window.

Note: nodeIntegration must be true and contextIsolation must be false in the webPreferences for this window

ipcRendererCallFirstListener(page, message, ...args) ⇒ Promise.<unknown>

Call just the first listener for a given ipcRenderer channel in a given window. UNLIKE MOST Electron ipcRenderer listeners, this function SHOULD return a value.

This function does not send data between main and renderer processes. It simply retrieves data from the renderer process.

Note: nodeIntegration must be true for this BrowserWindow.

ipcRendererEmit(page, message, ...args) ⇒ Promise.<boolean>

Emit an IPC message to a given window. This will trigger all ipcRenderer listeners for the message.

This does not transfer data between main and renderer processes. It simply emits an event in the renderer process.

Note: nodeIntegration must be true for this window

clickMenuItemById(electronApp, id) ⇒ Promise.<void>

Execute the .click() method on the element with the given id. NOTE: All menu testing functions will only work with items in the application menu.

clickMenuItem(electronApp, property, value) ⇒ Promise.<void>

Click the first matching menu item by any of its properties. This is useful for menu items that don't have an id. HOWEVER, this is not as fast or reliable as using clickMenuItemById() if the menu item has an id.

NOTE: All menu testing functions will only work with items in the application menu.

getMenuItemAttribute(electronApp, menuId, attribute) ⇒ Promise.<string>

Get a given attribute the MenuItem with the given id.

getMenuItemById(electronApp, menuId) ⇒ Promise.<MenuItemPartial>

Get information about the MenuItem with the given id

getApplicationMenu(electronApp) ⇒ Promise.<Array.<MenuItemPartial>>

Get the current state of the application menu. Contains only primitive values and submenus.. Very similar to menu construction template structure in Electron.

findMenuItem(electronApp, property, value, menuItems) ⇒ Promise.<MenuItemPartial>

Find a MenuItem by any of its properties

waitForMenuItem(electronApp, id) ⇒ Promise.<void>

Wait for a MenuItem to exist

waitForMenuItemStatus(electronApp, id, property, value) ⇒ Promise.<void>

Wait for a MenuItem to have a specific attribute value. For example, wait for a MenuItem to be enabled... or be visible.. etc

addTimeoutToPromise(promise, timeoutMs, timeoutMessage) ⇒ Promise.<T>

Add a timeout to any Promise

addTimeout(functionName, timeoutMs, timeoutMessage, ...args) ⇒ Promise.<T>

Add a timeout to any helper function from this library which returns a Promise.

Typedefs

ElectronAppInfo

Format of the data returned from parseElectronApp()

findLatestBuild(buildDirectory) ⇒ string

Parses the out directory to find the latest build of your Electron project. Use npm run package (or similar) to build your app prior to testing.

Assumptions: We assume that your build will be in the out directory, and that the build directory will be named with a hyphen-delimited platform name, e.g. out/my-app-win-x64. If your build directory is not out, you can pass the name of the directory as the buildDirectory parameter. If your build directory is not named with a hyphen-delimited platform name, this function will not work. However, you can pass the build path into parseElectronApp() directly.

Kind: global function
Returns: string -

• path to the most recently modified build directory

parseElectronApp(buildDir) ⇒ ElectronAppInfo

Given a directory containing an Electron app build, or the path to the app itself (directory on Mac, executable on Windows), return a bunch of metadata, including the path to the app's executable and the path to the app's main file.

Format of the data returned is an object with the following properties:

• executable: path to the app's executable file
• main: path to the app's main (JS) file
• name: name of the app
• resourcesDir: path to the app's resources directory
• asar: true if the app is using asar
• platform: OS platform
• arch: architecture
• packageJson: the JSON.parse()'d contents of the package.json file.

Kind: global function
Returns: ElectronAppInfo -

metadata about the app

electronWaitForFunction(electronApp, fn, arg) ⇒ Promise.<void>

Wait for a function to evaluate to true in the main Electron process. This really should be part of the Playwright API, but it's not.

This function is to electronApp.evaluate() as page.waitForFunction() is page.evaluate().

Kind: global function
Fulfil: void Resolves when the function returns true

ElectronAppInfo

Format of the data returned from parseElectronApp()

Kind: global typedef
Properties

stubDialog(app, method, value) ⇒ Promise.<void>

Stub a single dialog method. This is a convenience function that calls stubMultipleDialogs for a single method.

Playwright does not have a way to interact with Electron dialog windows, so this function allows you to substitute the dialog module's methods during your tests. By stubbing the dialog module, your Electron application will not display any dialog windows, and you can control the return value of the dialog methods. You're basically saying "when my application calls dialog.showOpenDialog, return this value instead". This allows you to test your application's behavior when the user selects a file, or cancels the dialog, etc.

Note: Each dialog method can only be stubbed with one value at a time, so you will want to call stubDialog before each time that you expect your application to call the dialog method.

Kind: global function
Returns: Promise.<void> -

A promise that resolves when the mock is applied.

Example

await stubDialog(app, 'showOpenDialog', {
 filePaths: ['/path/to/file'],
 canceled: false,
})
await clickMenuItemById(app, 'open-file')
// when time your application calls dialog.showOpenDialog,
// it will return the value you specified

stubMultipleDialogs(app, mocks) ⇒ Promise.<void>

Stub methods of the Electron dialog module.

Playwright does not have a way to interact with Electron dialog windows, so this function allows you to mock the dialog module's methods during your tests. By mocking the dialog module, your Electron application will not display any dialog windows, and you can control the return value of the dialog methods. You're basically saying "when my application calls dialog.showOpenDialog, return this value instead". This allows you to test your application's behavior when the user selects a file, or cancels the dialog, etc.

Kind: global function
Returns: Promise.<void> -

A promise that resolves when the mocks are applied.

Example

await stubMultipleDialogs(app, [
 {
   method: 'showOpenDialog',
   value: {
     filePaths: ['/path/to/file1', '/path/to/file2'],
     canceled: false,
   },
 },
 {
    method: 'showSaveDialog',
    value: {
      filePath: '/path/to/file',
      canceled: false,
    },
  },
])
await clickMenuItemById(app, 'save-file')
// when your application calls dialog.showSaveDialog,
// it will return the value you specified

stubAllDialogs(app) ⇒ Promise.<void>

Stub all dialog methods. This is a convenience function that calls stubMultipleDialogs for all dialog methods. This is useful if you want to ensure that dialogs are not displayed during your tests. However, you may want to use stubDialog or stubMultipleDialogs to control the return value of specific dialog methods (e.g. showOpenDialog) during your tests.

Kind: global function
Returns: Promise.<void> -

A promise that resolves when the mocks are applied.

ipcMainEmit(electronApp, message, ...args) ⇒ Promise.<boolean>

Emit an ipcMain message from the main process. This will trigger all ipcMain listeners for the message.

This does not transfer data between main and renderer processes. It simply emits an event in the main process.

Kind: global function
Category: IPCMain
Fulfil: boolean true if there were listeners for this message
Reject: Error if there are no ipcMain listeners for the event

ipcMainCallFirstListener(electronApp, message, ...args) ⇒ Promise.<unknown>

Call the first listener for a given ipcMain message in the main process and return its result.

NOTE: ipcMain listeners usually don't return a value, but we're using this to retrieve test data from the main process.

Generally, it's probably better to use ipcMainInvokeHandler() instead.

Kind: global function
Category: IPCMain
Fulfil: unknown resolves with the result of the function
Reject: Error if there are no ipcMain listeners for the event

ipcMainInvokeHandler(electronApp, message, ...args) ⇒ Promise.<unknown>

Get the return value of an ipcMain.handle() function

Kind: global function
Category: IPCMain
Fulfil: unknown resolves with the result of the function called in main process

ipcRendererSend(page, channel, ...args) ⇒ Promise.<unknown>

Send an ipcRenderer.send() (to main process) from a given window.

Note: nodeIntegration must be true and contextIsolation must be false in the webPreferences for this BrowserWindow.

Kind: global function
Category: IPCRenderer
Fulfil: unknown resolves with the result of ipcRenderer.send()

ipcRendererInvoke(page, message, ...args) ⇒ Promise.<unknown>

Send an ipcRenderer.invoke() from a given window.

Note: nodeIntegration must be true and contextIsolation must be false in the webPreferences for this window

Kind: global function
Category: IPCRenderer
Fulfil: unknown resolves with the result of ipcRenderer.invoke()

ipcRendererCallFirstListener(page, message, ...args) ⇒ Promise.<unknown>

Call just the first listener for a given ipcRenderer channel in a given window. UNLIKE MOST Electron ipcRenderer listeners, this function SHOULD return a value.

This function does not send data between main and renderer processes. It simply retrieves data from the renderer process.

Note: nodeIntegration must be true for this BrowserWindow.

Kind: global function
Category: IPCRenderer
Fulfil: unknown the result of the first ipcRenderer.on() listener

ipcRendererEmit(page, message, ...args) ⇒ Promise.<boolean>

Emit an IPC message to a given window. This will trigger all ipcRenderer listeners for the message.

This does not transfer data between main and renderer processes. It simply emits an event in the renderer process.

Note: nodeIntegration must be true for this window

Kind: global function
Category: IPCRenderer
Fulfil: boolean true if the event was emitted
Reject: Error if there are no ipcRenderer listeners for the event

clickMenuItemById(electronApp, id) ⇒ Promise.<void>

Execute the .click() method on the element with the given id. NOTE: All menu testing functions will only work with items in the application menu.

Kind: global function
Category: Menu
Fulfil: void resolves with the result of the click() method - probably undefined

clickMenuItem(electronApp, property, value) ⇒ Promise.<void>

Click the first matching menu item by any of its properties. This is useful for menu items that don't have an id. HOWEVER, this is not as fast or reliable as using clickMenuItemById() if the menu item has an id.

NOTE: All menu testing functions will only work with items in the application menu.

Kind: global function
Category: Menu
Fulfil: void resolves with the result of the click() method - probably undefined

getMenuItemAttribute(electronApp, menuId, attribute) ⇒ Promise.<string>

Get a given attribute the MenuItem with the given id.

Kind: global function
Category: Menu
Fulfil: string resolves with the attribute value

getMenuItemById(electronApp, menuId) ⇒ Promise.<MenuItemPartial>

Get information about the MenuItem with the given id

Kind: global function
Category: Menu
Fulfil: MenuItemPartial the MenuItem with the given id

getApplicationMenu(electronApp) ⇒ Promise.<Array.<MenuItemPartial>>

Get the current state of the application menu. Contains only primitive values and submenus.. Very similar to menu construction template structure in Electron.

Kind: global function
Category: Menu
Fulfil: MenuItemPartial[] an array of MenuItem-like objects

findMenuItem(electronApp, property, value, menuItems) ⇒ Promise.<MenuItemPartial>

Find a MenuItem by any of its properties

Kind: global function
Category: Menu
Fulfil: MenuItemPartial the first MenuItem with the given property and value

waitForMenuItem(electronApp, id) ⇒ Promise.<void>

Wait for a MenuItem to exist

Kind: global function
Category: Menu
Fulfil: void resolves when the MenuItem is found

waitForMenuItemStatus(electronApp, id, property, value) ⇒ Promise.<void>

Wait for a MenuItem to have a specific attribute value. For example, wait for a MenuItem to be enabled... or be visible.. etc

Kind: global function
Category: Menu
Fulfil: void resolves when the MenuItem with correct status is found

addTimeoutToPromise(promise, timeoutMs, timeoutMessage) ⇒ Promise.<T>

Add a timeout to any Promise

Kind: global function
Returns: Promise.<T> -

the result of the original promise if it resolves before the timeout

addTimeout(functionName, timeoutMs, timeoutMessage, ...args) ⇒ Promise.<T>

Add a timeout to any helper function from this library which returns a Promise.

Kind: global function
Returns: Promise.<T> -

the result of the helper function if it resolves before the timeout